In what follows the text has been generated by a person (me – Rod Rivers) while the graphics are mainly AI.
Introduction
Will robots have the good manners not to wipe out humanity? As these systems rapidly advance in autonomy and intelligence, the danger seems increasingly real. Are generative AI systems like Open AI’s Chat GPT, Google’s Gemini, and Anthropic’s Claud a good basis or model for the future.
Even if they can do many things better than we can, will they ever understand us? What control mechanisms do they need to mitigate their potential dangers and how can we stop power being concentrated in a small number of companies that win the race to Artificial General Intelligence (AGI)? Are we even thinking about AI systems the right way or is our thinking rooted in outdated conceptions of AI and society?
While Large Language Models are enormously useful in answering questions etc., we have been going wrong with most other applications of artificial intelligence (AI). AIs have such bad manners. They hog our attention. They can tell us things that aren’t true. They help dubious people invent and widely disseminate fake news so we no longer know what’s true and what isn’t. They influence us to vote for questionable people and policies. They suggest we buy things we never knew we wanted, and they steal all our private data and sell it to the highest bidder. Then they take all our jobs. That’s altogether very bad manners. What’s more, current AI may seem intelligent, but by human standards it’s rather stupid. And already, even though people are still in the loop, AI is controlling lethal weapons and gradually taking over the running of the world. That’s not just bad manners, it’s dangerous.
Another big question – should we build AI systems to think and act like people – in our own image, so to speak? Individual humans are fallible. We have all kinds of limitations, prejudices, and biases. We live in our own bubbles of emotion, evidence and reason based on the circumstances of our birth and the history of our experience. It’s hard to see outside of our own point of view and consider multiple perspectives. It’s also hard to see the big picture of how people and robots might interact and exist co-operatively together in society. Driven by profit and cost-saving, we are currently implementing AIs to duplicate our previous systems, locking-in the ways of working, prejudices and biases built into our organisational decision-making. Should AIs and robots of the future be like us, with the possibility that by replicating human characteristics they may gradually replace us and take away our purpose? On the other hand, if we don’t build them like humans will we ever understand them.
The average age of a European citizen is about 40 years. If you are under the age of 20 you can expect to live most of your life along-side intelligent machines. If you are 60 or even 70, there is still a very good chance of seeing Artificial General Intelligence (AGI). There is a brief window during the next few years when, having had a glimpse of what might be possible with AI, we can take stock. We can stand to one side of technology speeding down its own track, and work from first principles of values and ethics to make another path. This is sometimes referred to as the alignment problem –how to align the goals of AI and robots with human goals and values. It is possible to build good manners and ethics into the detailed working of AI and robots. But what good manners and whose ethics? Should AI have fixed ethical principles, or will it be possible to train AIs with the ethics of our own choosing, and how would this work?
This book explains how AI with manners might come about. It sets out what may give autonomous intelligent machines the capability to regulate their own actions so that they add to human wellbeing and minimise the possibility of harm. The many advances in our scientific understanding and our improving technological capabilities, combined with our insatiable curiosity and desire to ‘make progress’, are all driving us towards an increasingly mechanised future. That is not necessarily a bad thing, but it might easily turn bad if we can neither understand nor control the artefacts we create.
AI can be positioned in many ways in society – as an oracle, a personal assistant, a benevolent master, a slave, and as an aid to government and organisations. AI has the potential to be a catalyst for social change. We are at an inflection point where these issues and ideas must be widely discussed. Like an infant, AI will grow up to adopt the values of those around it. Society needs to be mature enough to teach it manners.
Good Manners for Robots is a book by Rod Rivers for release in Spring 2026. Email rod.rivers@deep-decisions.com with subject ‘Preview‘ to receive a free .pdf preview of part 1.